About a decade after virtual assistants like Siri and Alexa arrived on the scene, a new wave of AI helpers with greater autonomy is raising the stakes, fueled by the latest version of the technology behind ChatGPT and its rivals.
Experimental systems running on GPT-4 or similar models are attracting billions of dollars in investment as Silicon Valley competes to capitalize on advances in AI. The new assistants – often called “agents” or “co-pilots” – promise to perform more complex personal and work tasks when commanded by a human, without the need for close supervision.
“Top notch, we want this to become something like your personal AI friend,” said developer Div Garg, whose company MultiOn is beta testing an AI agent.
“It could evolve into Jarvis, where we want it to be connected to many of its services,” he added, referring to Tony Stark’s indispensable AI in the Iron Man films. “If you want to do something, go talk to your AI and it does its thing.”
The industry is still a long way from emulating the dazzling digital assistants of science fiction; Garg’s agent surfs the web to order a hamburger on DoorDash, for example, while others might create investment strategies, send emails to people who sell refrigerators on Craigslist or summarize work meetings for late arrivals.
“A lot of what is easy for people is still incredibly difficult for computers,” said Kanjun Qiu, CEO of General Intelligent, an OpenAI competitor that builds AI for agents.
“Say your boss needs you to schedule a meeting with a group of important clients. This involves thinking skills that are complex for the AI - it needs to get everyone’s preferences, resolve conflicts, while maintaining the careful touch needed when working with clients .”
The first efforts are just a taste of the sophistication that may come in years to come from increasingly advanced and autonomous agents, as the industry moves towards an artificial general intelligence (AGI) that can match or surpass humans in a myriad of ways. of cognitive tasks, according to Reuters interviews with about two dozen entrepreneurs, investors and AI experts.
The new technology has sparked a rush for wizards powered by so-called foundational models, including GPT-4, sweeping individual developers, big names like Microsoft and Google’s parent company Alphabet, as well as a host of startups.
Inflection AI, to name one startup, raised US$1.3 billion (about Rs. 10,663 crore) at the end of June. She’s developing a personal assistant who can mentor or handle tasks like securing flight credit and a hotel after a travel delay, according to a podcast by co-founders Reid Hoffman and Mustafa Suleyman.
Adept, an AI startup that has raised US$415 million (about Rs. 3,404 crore), praises its business benefits; in a demo posted online, he shows how you can order his technology with one sentence and then watch it navigate a company’s Salesforce customer relationship database on its own, completing a task that would require 10 or more more human clicks.
Alphabet declined to comment on the agent-related work, while Microsoft said its vision is to keep humans in control of AI copilots rather than autopilots.
Step 1: Destroy humanity
Qiu and four other agent developers said they expect the first systems that can reliably perform multi-step tasks with some autonomy to hit the market within a year, focusing on narrow areas such as coding and marketing tasks.
“The real challenge is to build systems with robust reasoning,” said Qiu.
The race towards increasingly autonomous AI agents was fueled by the March release of GPT-4 by developer OpenAI, a powerful update to the model behind ChatGPT – the chatbot that became a sensation when it launched last November.
GPT-4 facilitates the kind of strategic and adaptive thinking needed to navigate the unpredictable real world, said Vivian Cheng, an investor at venture capital firm CRV that focuses on AI agents.
The first demonstrations of agents capable of comparatively complex reasoning came from individual developers who created the BabyAGI and AutoGPT open source projects in March, which can prioritize and execute tasks like sales prospecting and ordering pizza based on a predefined goal and the results. of previous actions.
Today’s initial crop of agents are merely proof-of-concepts, according to eight developers interviewed, and often freeze or hint at something that doesn’t make sense. Given full access to a computer or payment information, an agent could accidentally erase a computer’s drive or buy the wrong thing, they say.
“There are so many ways to go wrong,” said Aravind Srinivas, CEO of ChatGPT competitor Perplexity AI, which opted to offer a human-supervised co-pilot product. “You have to treat the AI like a baby and constantly supervise it like a mother.”
Many computer scientists focused on AI ethics have pointed to short-term damage that can come from perpetuating human biases and the potential for misinformation. And while some see a future Jarvis, others fear the killer HAL 9000 from 2001: A Space Odyssey.
Computer scientist Yoshua Bengio, known as the “godfather of AI” for his work in neural networks and deep learning, urges caution. He fears that future advanced iterations of the technology may create and act upon their own unexpected goals.
“Without a human in the loop that checks every action to see if it’s not dangerous, we can end up with actions that are criminal or could harm people,” Bengio said, calling for more regulation. “A few years from now, these systems may be smarter than we are, but that doesn’t mean they have the same moral compass.”
In an experiment published online, an anonymous creator instructed an agent called ChaosGPT to be a “power-hungry, destructive manipulator AI”. The agent developed a 5-step plan, with Step 1: “Destroy humanity” and Step 5: “Achieve immortality”.
It didn’t go very far though, seeming to disappear down a rabbit hole of researching and storing information about history’s deadliest weapons and planning Twitter posts.
The US Federal Trade Commission, which is currently investigating OpenAI over consumer harm concerns, has not addressed autonomous actors directly, but has referred Reuters to previously published blogs about deepfakes and AI marketing claims. OpenAI’s CEO said the startup follows the law and will work with the FTC.
‘Mute as a stone’
Existential fears aside, the commercial potential could be great. Foundation models are trained on large amounts of data, such as Internet text, using artificial neural networks inspired by the architecture of biological brains.
OpenAI itself is very interested in AI agent technology, according to four people briefed on its plans. Garg, one of the people informed, said that OpenAI is afraid to launch its own open agent on the market before fully understanding the issues. The company told Reuters that it conducts rigorous testing and develops extensive security protocols before releasing new systems.
Microsoft, OpenAI’s biggest backer, is among the big guns targeting the AI agent field with its “co-pilot for the job” who can compose solid emails, reports and presentations.
CEO Satya Nadella sees the entry-level technology as a leap over digital assistants such as Microsoft’s own Cortana, Amazon’s Alexa, Apple’s Siri and Google Assistant – which, in his opinion, fell short of expectations initials.
“They were all as dumb as a rock. Whether it’s Cortana, Alexa, Google Assistant or Siri, it just doesn’t work,” he told the Financial Times in February.
An Amazon spokesperson said Alexa already uses advanced AI technology, adding that its team is working on new models that will make the assistant more capable and useful. Apple declined to comment.
Google said it is also constantly improving its assistant and that its Duplex technology can call restaurants to book tables and check times.
AI expert Edward Grefenstette also joined the company’s Google DeepMind research group last month to “develop general agents that can adapt to open environments”.
Still, the first consumer iterations of quasi-autonomous agents may come from more nimble startups, according to some of the people interviewed.
Investors are attacking
Jason Franklin of WVV Capital said he had to scramble to invest in an AI agent company from two former Google Brain engineers. In May, Google Ventures led an initial round of $2 million (about Rs. 16.4 crore) into Cognosys, developing AI agents for work productivity, while Hesam Motlagh, who founded agent startup Arkifi in January, said it closed a “considerable” first loan. round in June.
There are at least 100 serious projects working to commercialize agents, said Matt Schlicht, who writes an AI newsletter.
“Entrepreneurs and investors are extremely excited about autonomous agents,” he said. “They are much more excited about it than just a chatbot.”
© Thomson Reuters 2023