Tools & Platforms
Generative AI vs. regenerative AI: Key differences explored
Like many types of technology, artificial intelligence isn’t a single, uniform entity. There are different types of AI, and each has a different way of working, a different purpose, and a different effect on business operations and processes.
Generative AI (GenAI) has become well-known in recent years and is now commonly found among all types of technology users. With GenAI, users can easily summarize, create text and images, and get knowledge-based responses to prompts.
Regenerative AI, however, is a lesser-known and emerging concept. Rather than focusing on creating new content, regenerative AI emphasizes continuous self-improvement, adaptability and autonomous system optimization.
What is generative AI?
GenAI is a type of AI that generates new content such as text, image, audio and video. That new content is derived from patterns that a GenAI model has learned from training data.
The training process involves self-supervised learning on millions to trillions of data points, enabling models to generate contextually relevant and creative outputs through natural-language interfaces. GenAI uses deep learning, generative adversarial networks (GANs) and transformer-based AI architectures such as large language models.
Key use cases for generative AI include the following:
- Content summarization. It can summarize different types of content.
- Text generation. It can write any type of text-based content — including articles, reports and marketing copy.
- Image and video creation. It can produce any type of image, as well as video content.
- Music and audio. It can compose original music and generate voiceovers.
- Chatbots and virtual assistants. It commonly powers chatbots and virtual assistants, providing users with natural language interfaces to access information.
- Code generation. It can assist developers with code suggestions and automate software development.
There is an ever-growing list of GenAI tools, including the following:
- ChatGPT. The most widely used GenAI tool, OpenAI’s ChatGPT provides a conversational AI interface for content generation and Q&A.
- Gemini. Google’s Gemini is an advanced family of multimodal AI models that helps users summarize and generate content.
- Google AI Overviews. The Google search engine integrates GenAI-powered technology to provide clear and succinct answers to user queries. These AI Overviews typically appear at the top of search results.
- Midjourney. While there is no shortage of text-to-image generation tools, one example is Midjourney, which lets users create any type of image from a simple text prompt.
- GitHub Copilot. GitHub Copilot provides AI-powered code completion and suggestions.
GenAI is having a widespread effect across multiple industries, including the following:
- Media and entertainment. GenAI creates content, composes music and assists with video production.
- Application development. AI-powered development tooling is making it easier to build applications.
- Healthcare. It supports drug discovery, medical imaging and personalized medicine.
- Finance. It automates reporting, fraud detection and customer service.
- E-commerce. It offers personalized marketing, product design and customer engagement.
What is regenerative AI?
Regenerative AI is an emerging area of AI development with models and platforms that regenerate — or self-repair — optimize and adapt over time. This is all done without any human intervention.
The basic idea is to mimic the ability of biological organisms to adapt to changes in the environment. With biological organisms, changes in response to various factors are sometimes a function of evolution. With technology, there is an attempt to follow the same process using evolutionary algorithms, which are a subset of evolutionary computation.
Regenerative AI also uses multiple techniques that somewhat mirror how humans learn and think. A couple of techniques include the following:
- Reinforcement learning. Reinforcement learning trains models to take desired actions by rewarding positive behaviors and punishing negative ones.
- Neuromorphic computing. Neuromorphic computing techniques are a core element of regenerative AI, providing mechanisms that attempt to work the same way as the human brain with neurons and synapses.
The effect of self-repair capabilities
The self-repair capability of regenerative AI is one of the most noteworthy aspects of the technology and has the potential for a significant effect on the AI-technology landscape.
Instead of requiring manual, human intervention to fix an issue or fine-tune and optimize, self-repair handles that automatically. It reduces or eliminates the need for hands-on human maintenance, which has the potential to be particularly valuable in remote or hazardous environments where human intervention is limited. Self-repair will also enhance overall AI system reliability, reduce downtime and reduce operational costs.
Regenerative AI has several capabilities, including the following:
- Self-repair. It can detect and fix errors or inefficiencies autonomously.
- Process optimization. It can identify and correct inefficient workflows.
- Continuous learning. It can adapt to new data and environments in real time.
- Fault tolerance. Thanks to self-repair, regenerative AI models are fault-tolerant.
While currently still in the early stages of development, regenerative AI has potential for a variety of applications, including the following:
- Robotics. It is ideal for robotics, where systems can self-diagnose and fix malfunctions.
- Autonomous vehicles. It could be used to help autonomous vehicles adapt to changing road conditions.
- Cybersecurity. Regenerative AI could be used to help counter new cyber threats in real time.
- Electricity distribution. It could power smart grids that dynamically optimize energy use.
- Remote locations. In remote locations where connectivity is limited, its ability to self-repair would be extremely useful.
Differences between generative and regenerative AI
While both generative and regenerative AI fall under the umbrella of artificial intelligence, they operate on different principles. The following table summarizes their key differences:
| Aspect | Generative AI | Regenerative AI |
| Definition | Generates new content based on training data. | Can self-repair, adapt and improve over time. |
| Core technology | Transformer-based neural networks, GANs and diffusion models. | Reinforcement learning, evolutionary algorithms and neuromorphic computing. |
| Learning approach | Static training on massive datasets with periodic fine-tuning. | Continuous learning through real-time feedback and experience. |
| Maintenance needs | Requires human intervention for updates and troubleshooting. | Self-maintains through autonomous error detection and correction. |
| Output focus | Creative content (text, images, code and audio). | System improvements and adaptive responses. |
| Market maturity | Wide commercial deployment in 2025. | Currently in experimental stage with limited practical applications. |
Future trends for generative and regenerative AI
There is much to look forward to for both generative AI and regenerative AI.
Trends show several future developments for generative AI, including the following:
- Agentic AI. GenAI is moving in a somewhat autonomous direction already with the growth of agentic AI, which can act and connect to different systems on behalf of users.
- Multimodal models. GenAI models are going multimodal, with single models able to understand and generate text, audio, images and video.
- Regulatory initiatives. There is a growing emphasis on addressing ethical concerns, such as user privacy and ensuring responsible use.
Regenerative AI also shows trends toward future developments, including the following:
- Transition from theoretical to practical. Regenerative AI has some ground to cover before it will be widely available and practical to deploy. In the coming years, the technology is expected to mature as computational hardware, software and algorithms improve.
- Advancements in neuromorphic computing. New forms of neuromorphic computing hardware, including silicon hardware, will be a key step in future development.
- Integration with the internet of things and edge computing. As the technology matures, it will find a natural fit in internet of things and edge computing deployments, providing the ability to self-optimize to changing conditions in real time.
Sean Michael Kerner is an IT consultant, technology enthusiast and tinkerer. He has pulled Token Ring, configured NetWare and been known to compile his own Linux kernel. He consults with industry and media organizations on technology issues.
Continue Reading
By JOCELYN GECKER
Associated Press
The book report is now a thing of the past. Take-home tests and essays are becoming obsolete.
Student use of artificial intelligence has become so prevalent, high school and college educators say, that to assign writing outside of the classroom is like asking students to cheat.
“The cheating is off the charts. It’s the worst I’ve seen in my entire career,” says Casey Cuny, who has taught English for 23 years. Educators are no longer wondering if students will outsource schoolwork to AI chatbots. “Anything you send home, you have to assume is being AI’ed.”
The question now is how schools can adapt, because many of the teaching and assessment tools that have been used for generations are no longer effective. As AI technology rapidly improves and becomes more entwined with daily life, it is transforming how students learn and study and how teachers teach, and it’s creating new confusion over what constitutes academic dishonesty.
“We have to ask ourselves, what is cheating?” says Cuny, a 2024 recipient of California’s Teacher of the Year award. “Because I think the lines are getting blurred.”
Cuny’s students at Valencia High School in southern California now do most writing in class. He monitors student laptop screens from his desktop, using software that lets him “lock down” their screens or block access to certain sites. He’s also integrating AI into his lessons and teaching students how to use AI as a study aid “to get kids learning with AI instead of cheating with AI.”
In rural Oregon, high school teacher Kelly Gibson has made a similar shift to in-class writing. She is also incorporating more verbal assessments to have students talk through their understanding of assigned reading.
“I used to give a writing prompt and say, ‘In two weeks, I want a five-paragraph essay,’” says Gibson. “These days, I can’t do that. That’s almost begging teenagers to cheat.”
Take, for example, a once typical high school English assignment: Write an essay that explains the relevance of social class in “The Great Gatsby.” Many students say their first instinct is now to ask ChatGPT for help “brainstorming.” Within seconds, ChatGPT yields a list of essay ideas, plus examples and quotes to back them up. The chatbot ends by asking if it can do more: “Would you like help writing any part of the essay? I can help you draft an introduction or outline a paragraph!”
Students are uncertain when AI usage is out of bounds
Students say they often turn to AI with good intentions for things like research, editing or help reading difficult texts. But AI offers unprecedented temptation, and it’s sometimes hard to know where to draw the line.
College sophomore Lily Brown, a psychology major at an East Coast liberal arts school, relies on ChatGPT to help outline essays because she struggles putting the pieces together herself. ChatGPT also helped her through a freshman philosophy class, where assigned reading “felt like a different language” until she read AI summaries of the texts.
“Sometimes I feel bad using ChatGPT to summarize reading, because I wonder, is this cheating? Is helping me form outlines cheating? If I write an essay in my own words and ask how to improve it, or when it starts to edit my essay, is that cheating?”
Her class syllabi say things like: “Don’t use AI to write essays and to form thoughts,” she says, but that leaves a lot of grey area. Students say they often shy away from asking teachers for clarity because admitting to any AI use could flag them as a cheater.
Schools tend to leave AI policies to teachers, which often means that rules vary widely within the same school. Some educators, for example, welcome the use of Grammarly.com, an AI-powered writing assistant, to check grammar. Others forbid it, noting the tool also offers to rewrite sentences.
“Whether you can use AI or not depends on each classroom. That can get confusing,” says Valencia 11th grader Jolie Lahey. She credits Cuny with teaching her sophomore English class a variety of AI skills like how to upload study guides to ChatGPT and have the chatbot quiz them, and then explain problems they got wrong.
But this year, her teachers have strict “No AI” policies. “It’s such a helpful tool. And if we’re not allowed to use it that just doesn’t make sense,” Lahey says. “It feels outdated.”
Schools are introducing guidelines, gradually
Many schools initially banned use of AI after ChatGPT launched in late 2022. But views on the role of artificial intelligence in education have shifted dramatically. The term “AI literacy” has become a buzzword of the back-to-school season, with a focus on how to balance the strengths of AI with its risks and challenges.
Over the summer, several colleges and universities convened their AI task forces to draft more detailed guidelines or provide faculty with new instructions.
The University of California, Berkeley emailed all faculty new AI guidance that instructs them to “include a clear statement on their syllabus about course expectations” around AI use. The guidance offered language for three sample syllabus statements — for courses that require AI, ban AI in and out of class, or allow some AI use.
“In the absence of such a statement, students may be more likely to use these technologies inappropriately,” the email said, stressing that AI is “creating new confusion about what might constitute legitimate methods for completing student work.”
Carnegie Mellon University has seen a huge uptick in academic responsibility violations due to AI, but often students aren’t aware they’ve done anything wrong, says Rebekah Fitzsimmons, chair of the AI faculty advising committee at the university’s Heinz College of Information Systems and Public Policy.
For example, one student who is learning English wrote an assignment in his native language and used DeepL, an AI-powered translation tool, to translate his work to English. But he didn’t realize the platform also altered his language, which was flagged by an AI detector.
Enforcing academic integrity policies has become more complicated, since use of AI is hard to spot and even harder to prove, Fitzsimmons said. Faculty are allowed flexibility when they believe a student has unintentionally crossed a line, but are now more hesitant to point out violations because they don’t want to accuse students unfairly. Students worry that if they are falsely accused, there is no way to prove their innocence.
Over the summer, Fitzsimmons helped draft detailed new guidelines for students and faculty that strive to create more clarity. Faculty have been told a blanket ban on AI “is not a viable policy” unless instructors make changes to the way they teach and assess students. A lot of faculty are doing away with take-home exams. Some have returned to pen and paper tests in class, she said, and others have moved to “flipped classrooms,” where homework is done in class.
Emily DeJeu, who teaches communication courses at Carnegie Mellon’s business school, has eliminated writing assignments as homework and replaced them with in-class quizzes done on laptops in “a lockdown browser” that blocks students from leaving the quiz screen.
“To expect an 18-year-old to exercise great discipline is unreasonable,” DeJeu said. “That’s why it’s up to instructors to put up guardrails.”
___
The Associated Press’ education coverage receives financial support from multiple private foundations. AP is solely responsible for all content. Find AP’s standards for working with philanthropies, a list of supporters and funded coverage areas at AP.org.
Tools & Platforms
Vibe coding has turned senior devs into ‘AI babysitters,’ but they say it’s worth it
Carla Rover once spent 30 minutes sobbing after having to restart a project she vibe coded.
Rover has been in the industry for 15 years, mainly working as a web developer. She’s now building a startup, alongside her son, that creates custom machine learning models for marketplaces.
She called vibe coding a beautiful, endless cocktail napkin on which one can perpetually sketch ideas. But dealing with AI-generated code that one hopes to use in production can be “worse than babysitting,” she said, as these AI models can mess up work in ways that are hard to predict.
She had turned to AI coding in a need for speed with her startup, as is the promise of AI tools.
“Because I needed to be quick and impressive, I took a shortcut and did not scan those files after the automated review,” she said. “When I did do it manually, I found so much wrong. When I used a third-party tool, I found more. And I learned my lesson.”
She and her son wound up restarting their whole project — hence the tears. “I handed it off like the copilot was an employee,” she said. “It isn’t.”
Rover is like many experienced programmers turning to AI for coding help. But such programmers are also finding themselves acting like AI babysitters — rewriting and fact-checking the code the AI spits out.
Techcrunch event
San Francisco
|
October 27-29, 2025
A recent report by content delivery platform company Fastly found that at least 95% of the nearly 800 developers it surveyed said they spend extra time fixing AI-generated code, with the load of such verification falling most heavily on the shoulders of senior developers.
These experienced coders have discovered issues with AI-generated code ranging from hallucinating package names to deleting important information and security risks. Left unchecked, AI code can leave a product far more buggy than what humans would produce.
Working with AI-generated code has become such a problem that it’s given rise to a new corporate coding job known as “vibe code cleanup specialist.”
TechCrunch spoke to experienced coders about their time using AI-generated code about what they see as the future of vibe coding. Thoughts varied, but one thing remained certain: The technology still has a long way to go.
“Using a coding co-pilot is kind of like giving a coffee pot to a smart six-year-old and saying, ‘Please take this into the dining room and pour coffee for the family,’” Rover said.
Can they do it? Possibly. Could they fail? Definitely. And most likely, if they do fail, they aren’t going to tell you. “It doesn’t make the kid less clever,” she continued. “It just means you can’t delegate [a task] like that completely.”
“You’re absolutely right!”
Feridoon Malekzadeh also compared vibe coding to a child.
He’s worked in the industry for more than 20 years, holding various roles in product development, software, and design. He’s building his own startup and heavily using vibe-coding platform Lovable, he said. For fun, he also vibe codes apps like one that generates Gen Alpha slang for Boomers.
He likes that he’s able to work alone on projects, saving time and money, but agrees that vibe coding is not like hiring an intern or a junior coder. Instead, vibe coding is akin to “hiring your stubborn, insolent teenager to help you do something,” he told TechCrunch.
“You have to ask them 15 times to do something,” he said. “In the end, they do some of what you asked, some stuff you didn’t ask for, and they break a bunch of things along the way.”
Malekzadeh estimates he spends around 50% of his time writing requirements, 10% to 20% of his time on vibe coding, and 30% to 40% of his time on vibe fixing — remedying the bugs and “unnecessary script” created by AI-written code.
He also doesn’t think vibe coding is the best at systems thinking — the process of seeing how a complex problem could impact an overall result. AI-generated code, he said, tries to solve more surface-level problems.
“If you’re creating a feature that should be broadly available in your product, a good engineer would create that once and make it available everywhere that it’s needed,” Malekzadeh said. “Vibe coding will create something five different times, five different ways, if it’s needed in five different places. It leads to a lot of confusion, not only for the user, but for the model.”
Meanwhile, Rover finds that AI “runs into a wall” when data conflicts with what it was hard-coded to do. “It can offer misleading advice, leave out key elements that are vital, or insert itself into a thought pathway you’re developing,” she said.
She also found that rather than admit to making errors, it will manufacture results.
She shared another example with TechCrunch, where she questioned the results an AI model initially gave her. The model started to give a detailed explanation pretending it used the data she uploaded. Only when she called it out did the AI model confess.
“It freaked me out because it sounded like a toxic co-worker,” she said.
On top of this, there are the security concerns.
Austin Spires is the senior director of developer enablement at Fastly and has been coding since the early 2000s.
He’s found through his own experience — along with chatting with customers — that vibe code likes to build what is quick rather than what is “right.” This may introduce vulnerabilities to the code of the kind that very new programmers tend to make, he said.
“What often happens is the engineer needs to review the code, correct the agent, and tell the agent that they made a mistake,” Spires told TechCrunch. “This pattern is why we’ve seen the trope of ‘you’re absolutely right’ appear over social media.”
He’s referring to how AI models, like Anthropic Claude, tend to respond “you’re absolutely right” when called out on their mistakes.
Mike Arrowsmith, the chief technology officer at the IT management software company NinjaOne, has been in software engineering and security for around 20 years. He said that vibe coding is creating a new generation of IT and security blind spots to which young startups in particular are susceptible.
“Vibe coding often bypasses the rigorous review processes that are foundational to traditional coding and crucial to catching vulnerabilities,” he told TechCrunch.
NinjaOne, he said, counters this by encouraging “safe vibe coding,” where approved AI tools have access controls, along with mandatory peer review and, of course, security scanning.
The new normal
While nearly everyone we spoke to agrees that AI-generated code and vibe-coding platforms are useful in many situations — like mocking up ideas — they all agree that human review is essential before building a business on it.
“That cocktail napkin is not a business model,” Rover said. “You have to balance the ease with insight.”
But for all the lamenting on its errors, vibe coding has changed the present and the future of the job.
Rover said vibe coding helped her tremendously in crafting a better user interface. Malekzadeh simply said that, despite the time he spends fixing code, he still gets more done with AI coders than without them.
“‘Every technology carries its own negativity, which is invented at the same time as technical progress,” Malekzadeh said, quoting the French theorist Paul Virilio, who spoke about inventing the shipwreck along with the ship.
The pros far outweigh the cons.
The Fastly survey found that senior developers were twice as likely to put AI-generated code into production compared to junior developers, saying that the technology helped them work faster.
Vibe coding is also part of Spires’ coding routine. He uses AI coding agents on several platforms for both front-end and back-end personal projects. He called the technology a mixed experience but said it’s good in helping with prototyping, building out boilerplate, or scaffolding out a test; it removes menial tasks so that engineers can focus on building, shipping, and scaling products.
It seems the extra hours spent combing through the vibe weeds will simply become a tolerated tax on using the innovation.
Elvis Kimara, a young engineer, is learning that now. He just graduated with a master’s in AI and is building an AI-powered marketplace.
Like many coders, he said vibe coding has made his job harder and has often found vibe coding a joyless experience.
“There’s no more dopamine from solving a problem by myself. The AI just figures it out,” he said. At one of his last jobs, he said senior developers didn’t look to help young coders as much — some not understanding new vibe-coding models, while others delegated mentorship tasks to said AI models.
But, he said, “the pros far outweigh the cons,” and he’s prepared to pay the innovation tax.
“We won’t just be writing code; we’ll be guiding AI systems, taking accountability when things break, and acting more like consultants to machines,” Kimara said of the new normal for which he’s preparing.
“Even as I grow into a senior role, I’ll keep using it,” he continued. “It’s been a real accelerator for me. I make sure I review every line of AI-generated code so I learn even faster from it.”
Over the past five years, the IDF has been working to transform itself into a network-enabled combat machine, with AI and Big Data enabling the flow of information across units and commands.
-
Business2 weeks agoThe Guardian view on Trump and the Fed: independence is no substitute for accountability | Editorial
-
Tools & Platforms1 month ago
Building Trust in Military AI Starts with Opening the Black Box – War on the Rocks
-
Ethics & Policy2 months agoSDAIA Supports Saudi Arabia’s Leadership in Shaping Global AI Ethics, Policy, and Research – وكالة الأنباء السعودية
-
Events & Conferences4 months agoJourney to 1000 models: Scaling Instagram’s recommendation system
-
Jobs & Careers3 months agoMumbai-based Perplexity Alternative Has 60k+ Users Without Funding
-
Podcasts & Talks2 months agoHappy 4th of July! 🎆 Made with Veo 3 in Gemini
-
Education3 months agoVEX Robotics launches AI-powered classroom robotics system
-
Education2 months agoMacron says UK and France have duty to tackle illegal migration ‘with humanity, solidarity and firmness’ – UK politics live | Politics
-
Funding & Business3 months agoKayak and Expedia race to build AI travel agents that turn social posts into itineraries
-
Podcasts & Talks2 months agoOpenAI 🤝 @teamganassi