The holiday shopping season has started fast, especially online, and those automated chatbots are working overtime.
Artificial intelligence is increasingly finding its way into decisions that once were purely human, like whether or not to buy a sweater. This year, more than ever, smart computer programs are stepping between customers and their shopping.
These programs make recommendations based on purchase history, browsing behavior and demographics. They also sharpen the results from online product searches, adjust prices based on competitive factors, and improve product placement or promotions. AI-powered customer service can answer questions and take orders. AI can even enable apparel shoppers to virtually “try on” clothes to see if they’ll fit.
This integration of AI with routine tasks of our everyday lives isn’t new, but its ubiquity is growing. Voice-operated digital assistants like Apple’s Siri, Amazon’s Alexa and Google Assistant remind us of appointments, tell us if it’s cold outside and control smart devices at home. The ChatGPT program and its imitators help draft emails and term papers. AI-powered streaming platforms like Netflix and music services like Spotify make uncanny recommendations, sometimes seeming to know our interests better than we know ourselves.
Merchants say their embrace of AI creates more efficient and personal shopping experiences. If Santa can tell when you’ve been bad or good, as the Christmas carol goes, imagine what a computer analyzing every keystroke can surmise about you and your holiday shopping list.
Some AI applications rival science fiction. A recent survey of 2,000 adults under age 40 showed that 13% of young men and 9% of young women are open to friendships with AI-generated companions, and one in four young people say they believe AI partners could eventually replace real-life romance. The Institute for Family Studies, a conservative think tank that conducted the poll, concludes, “Robots aren’t just coming for your jobs but for your relationships, too.”
In the U.S., this incursion is largely unregulated, which could be risky based on how AI is being used. The European Union has developed a sensible AI regulatory framework that rates the risk from “minimal” to “unacceptable.”
Minimal risks include spam filters that use AI to better screen out junk emails and the like. An unacceptable risk threatens fundamental rights, such as using facial recognition and deepfakes to influence elections. Under the EU system, that’s not allowed.
“High-risk” AI covers a lot of sensitive ground, from managing road traffic to grading students and evaluating employees. The risk comes from the potential for things to go badly wrong, and the EU requires those systems to meet stringent standards before being unleashed on the public.
Most AI related to holiday shopping falls into the catch-all category of “limited” risk, meaning the systems could be used to deceive people in relatively small ways. For example, the EU requires that chatbot programs conversing in text or voice make it clear that human users are interacting with AI, not other humans.
The EU framework is still being sorted out, but at least the Europeans have one. In the U.S., Congress so far has failed to pass legislation, and the only national standards to date stemmed from an executive order that outgoing President Joe Biden imposed in 2023.
With ex-President Donald Trump returning to the White House, Biden’s order is probably headed for the trash, and it’s unclear what AI guardrails a Trump administration might want, if any. During the run-up to Election Day, both Republicans and Democrats used AI to target messages and generate memes. Fortunately, the worst fears of a lifelike fake message or video disrupting the election never materialized.
With the GOP in charge of Congress, there may be an appetite for AI legislation to protect national security or ban the awful practice of creating non-consensual explicit images. But any effort to curb the spread of misinformation faces a tough path as Republican lawmakers resist anything they see as suppressing free speech.
The problem is that with no safeguards, Americans are likely to be confronted with AI-generated content that isn’t easily recognized as AI-generated. In that instance, people could lose confidence in their ability to determine what’s true and what’s being foisted on them for commercial purposes, or worse. Holiday shopping is just one example of a seemingly benign activity that could become problematic if AI is taken too far.
Advocates in the AI industry have pushed for rules of the road to help the technology gain acceptance and head off problems. The Europeans have provided a reasonable starting point. Now it’s up to the ruling GOP to recognize that imposing regulations on the AI free-for-all stands to promote innovation and ensure that these powerful tools are used in a manner that serves the public.
Submit a letter, of no more than 400 words, to the editor here or email letters@chicagotribune.com.