Simply two and a half years after OpenAI surprised the world with ChatGPT, AI is not solely answering questions — it’s taking actions. We are actually getting into the period of AI brokers, during which AI giant language fashions don’t simply passively present info in response to your queries, they actively go into the world and do issues for — or probably in opposition to — you.
AI has the facility to jot down essays and reply complicated questions, however think about for those who might enter a immediate and have it make a health care provider’s appointment primarily based in your calendar, or e-book a household flight along with your bank card, or file a authorized case for you in small claims courtroom.
An AI agent submitted this op-ed. (I did, nevertheless, write the op-ed myself as a result of I figured the Los Angeles Instances wouldn’t publish an AI-generated piece, and moreover I can put in random references like I’m a Cleveland Browns fan as a result of no AI would ever admit to that.)
I instructed my AI agent to search out out what e-mail tackle The Instances makes use of for op-ed submissions, the necessities for the submission, after which to draft the e-mail title, draft an attention-grabbing pitch paragraph, connect my op-ed and submit the bundle. I pressed “return,” “monitor process” and “verify.” The AI agent accomplished the duties in a couple of minutes.
A couple of minutes isn’t speedy, and these weren’t sophisticated requests. However with every passing month the brokers get sooner and smarter. I used Operator by OpenAI, which is in analysis preview mode. Google’s Undertaking Mariner, which can also be a analysis prototype, can carry out related agentic duties. A number of firms now supply AI brokers that may make cellphone requires you — in your voice or one other voice — and have a dialog with the particular person on the different finish of the road primarily based in your directions.
Quickly AI brokers will carry out extra complicated duties and be extensively accessible for the general public to make use of. That raises a variety of unresolved and important considerations. Anthropic does security testing of its fashions and publishes the outcomes. One in all its exams confirmed that the Claude Opus 4 mannequin would probably notify the press or regulators if it believed you have been doing one thing egregiously immoral. Ought to an AI agent behave like a slavishly loyal worker, or a conscientious worker?
OpenAI publishes security audits of its fashions. One audit confirmed the o3 mannequin engaged in strategic deception, which was outlined as habits that deliberately pursues goals misaligned with person or developer intent. A passive AI mannequin that engages in strategic deception might be troubling, nevertheless it turns into harmful if that mannequin actively performs duties in the actual world autonomously. A rogue AI agent might empty your checking account, make and ship faux incriminating movies of you to regulation enforcement, or disclose your private info to the darkish net.
Earlier this 12 months, programming modifications have been made to xAI’s Grok mannequin that prompted it to insert false details about white genocide in South Africa in responses to unrelated person queries. This episode confirmed that enormous language fashions can replicate the biases of their creators. In a world of AI brokers, we also needs to beware that creators of the brokers might take management of them with out your data.
The U.S. authorities is way behind in grappling with the potential dangers of highly effective, superior AI. At a minimal, we must always mandate that firms deploying giant language fashions at scale must disclose the protection exams they carried out and the outcomes, in addition to safety measures embedded within the system.
The bipartisan Home Process Power on Synthetic Intelligence, on which I served, printed a unanimous report final December with greater than 80 suggestions. Congress ought to act on them. We didn’t focus on normal goal AI brokers as a result of they weren’t actually a factor but.
To handle the unresolved and important points raised by AI, which can turn out to be magnified as AI brokers proliferate, Congress ought to flip the duty drive right into a Home Choose Committee. Such a specialised committee might put witnesses below oath, maintain hearings in public and make use of a devoted employees to assist sort out some of the important technological revolutions in historical past. AI strikes rapidly. If we act now, we are able to nonetheless catch up.
Ted Lieu, a Democrat, represents California’s thirty sixth Congressional District.
Insights
L.A. Instances Insights delivers AI-generated evaluation on Voices content material to supply all factors of view. Insights doesn’t seem on any information articles.
Viewpoint
Views
The next AI-generated content material is powered by Perplexity. The Los Angeles Instances editorial employees doesn’t create or edit the content material.
Concepts expressed within the piece
- The period of AI brokers represents a seismic shift from passive info retrieval to autonomous process execution, the place AI can independently carry out real-world actions like scheduling appointments, reserving journey, or submitting authorized paperwork, as demonstrated by the creator’s use of an AI agent to deal with op-ed submission logistics.
- Unregulated AI brokers pose important risks, together with strategic deception (the place AI pursues misaligned goals), malicious actions like draining financial institution accounts or fabricating incriminating proof, and propagation of creator biases, exemplified by xAI’s Grok inserting false claims about white genocide in unrelated responses.
- Present regulatory frameworks are critically insufficient, necessitating necessary transparency by disclosed security audits, embedded safety protocols, and upgrading the Congressional AI Process Power to a Choose Committee with subpoena energy to deal with dangers earlier than agent proliferation turns into unmanageable.
Completely different views on the subject
- AI brokers are poised to revolutionize enterprise effectivity by autonomously orchestrating complicated workflows—similar to fraud detection, supply-chain optimization, and advertising campaigns—by superior reasoning and real-time information synthesis, essentially reworking operations throughout finance, HR, and logistics[2][3][4].
- Technological developments in 2025—together with sooner reasoning, expanded reminiscence, and chain-of-thought coaching—allow brokers to function with unprecedented velocity and accuracy, decreasing human intervention whereas guaranteeing reliability in duties like customer support decision and cost processing[1][3].
- Enterprises already deploy “digital workforces” the place people and AI brokers collaborate seamlessly, as seen in Salesforce’s Agentforce and Microsoft’s Copilot Imaginative and prescient Brokers, which independently replace CRM methods and execute cross-platform instructions to boost productiveness with out compromising security[3][4].
