اردو 
The Potential Perils of AI: Insights from Yoshua Bengio
The technology could become more intelligent than people and eventually take over, according to Yoshua Bengio
by Brian Contreras,
Los Angeles Times
But whatever happens, there's still 12 months left until the end of Hollywood's union pushback against automation. the rise of consumer chatbots, including OpenAI's GPT-4 and Elon Musk's Grok; Strike Back Against Sam Altman; initial doubts about the actions of the regulatory authorities; And of course, there's the viral deepfake of Pope Francis in a padded jacket.
The Times asked several experts and stakeholders for their AI predictions for 2024 to find out what we can expect in the new year. The results alternated between excitement, curiosity, and skepticism. It's a good mix of feelings about a technology that remains polarized and unpredictable. Regulators are stepping in, but not everyone is happy about it.
When a surgeon or a stockbroker goes to work, they do so with the support of a license or certification. Could 2024 be the year AI meets the same standards?
"Next year we may need professional licenses for AI systems," said Amy Webb, CEO of the Future Today Institute, a consulting firm. “Although some fields require specialized certification in humans, algorithms can work without standardized tests. Ideally, you wouldn't want to see a urologist without a doctor's license to perform the surgery.”
That would be a development that follows policy changes in recent months that have included good-faith efforts to regulate this powerful new technology, including President Biden's sweeping executive order and a Senate policy bill to crack down on fakes.
"I am deeply concerned about the potential impact [generative AI] could have on our democracy and institutions before the November election," said Sen. Chris Coons (D-Delaware), who sponsored the deepfake design. For next year. "Creators, experts, and the public are demanding federal protections to set clear policies on the use of generative AI, and Congress must do so."
Regulation is not just a domestic issue. Justin Hughes, a professor of commercial and intellectual property law at Loyola Law School, said he expects the European Union to finalize its AI legislation next year, leaving EU-wide AI regulation 24 months away.
Hughes said these include transparency and governance requirements, as well as banning dangerous uses of artificial intelligence, such as identifying someone's race and sexual orientation or manipulating their behavior. Like many European regulations, it could also affect US business. But as demands for a fence grew, a backlash emerged. In particular, effective acceleration (the so-called “e/acc” movement) has gained momentum through the demand for rapid innovation with limited political oversight.
Julie Fredrickson, a tech investor who joined the e/acc movement, said she expects regulatory tensions to intensify further in the new year.
"The biggest challenge we face is using [a tool] that counts the word ISIS to raise important constitutional questions for the United States that any regulatory framework must address," Fredrickson said. "The public must make it clear to the government that our fundamental rights, such as freedom of expression, cannot be compromised."
Authenticity will be more important than ever.
Imagine being able to know for sure if the vacation photo your friend just posted on Instagram is real or created on a server farm somewhere. Mike Gioia, co-founder of AI workflow startup Pickaxe, believes this will soon be possible.
In particular, he predicted that Apple will release an "iPhone photo" stamp next year that authenticates photos without AI. Other experts agree that efforts to build trust and authenticity will become more important as AI floods the Internet with synthetic text, photos, and videos (not to mention bots impersonating real people).
Andy Parsons, senior director of the Adobe Content Authenticity Initiative, said that "content credentials," or metadata contained in digital media files, determine who created something and with what tools, like a food label. He said he expected it to be recorded. These disruptive decisions could be especially important as the United States enters a presidential election year.
This is the first wave of cheap AI media coverage in history.
Bill Burton, a former Obama administration spokesman, predicted that "the most watched and most engaged videos in the 2024 election will be created by AI."
The steam engine of innovation continues to run…
The past year has seen significant advances in AI technology, and AI research and development continues to advance with major products, including ChatGPT, the fastest-growing consumer app in history, now in its fourth version.
Many AI insiders believe the pace of innovation will continue into the new year.
"Every user of enterprise and consumer applications uses AI and doesn't know it," said Ted Ross, CEO of the Los Angeles Information Technology Agency.
"We expect AI capabilities and highly visible [generative] AI platforms like ChatGPT to be quickly integrated into enterprise and consumer applications without users even realizing it."
Other developments may be more niche, but no less impactful. Some experts predict a proliferation of simpler, more focused alternatives to "large language models" based on ChatGPT and Grok. AI can also become more adept at improving itself.
"There aren't many tools to accelerate AI research," says Anastasis Germanidis, chief technology officer at synthetic video startup Runway. "A lot of these tools will probably come out in the next year" to help you write or debug code.
… As long as the bubble doesn't burst
The AI market is booming right now, but not everyone thinks the glory days will last. "Ashagan AI will either go bankrupt in 2024 or be acquired at an absurdly low price," recently wrote Clement Delangou, CEO of Hugging Face, an open-source AI development community.
Eric Siegel, former Columbia University professor and author of The AI Playbook: Mastering Rare Skills with Machine Learning, is more cautious. "There is a growing concern as it becomes increasingly clear that there is no such thing as a perfect [generative] AI application," Siegel told The Times, citing applications that encourage widespread use of AI.
"If today's high expectations are not met, despair will follow."
He warned that we may eventually be entering an "AI winter," a period of declining interest in (and investment in) the technology. But, he added, "'Fashion' has tremendous momentum right now, and that momentum will only increase as new, interesting and potentially valuable features continue to emerge."
Even skeptics seem to be expecting a bumper year for AI.
