Few technologies have sparked the human imagination like artificial intelligence has. Who among us couldn’t, on demand, conjure up a future scenario made possible through AI?

The fact is, the technology has developed to the point that we can, with a fair degree of certainty, predict actual AI-driven events that will happen in the next 12 months. Like most things in life, these events will be as much good as bad. But – call me an eternal AI optimist –if we use the less positive occurrences as fodder for corrective measures, it will only lead to better things ahead.  

Here are six predictions for AI in 2020, equal parts positive and negative. First the bad:

  • AI-generated content will influence an election. When OpenAI released its AI-based text generator GPT-2 in early 2019, it held back on a more robust version due to the company’s stated concerns about malicious uses of the technology. In addition to the hoped-for applications (AI writing assistants, unsupervised translation between languages), OpenAI acknowledged the potential to generate misleading news articles, impersonate others online and automate production of fake content on social media.

    By the end of the year, OpenAI went ahead with the full release, with the assurance that its output was only marginally more credible than the previous release from a human perspective. It also acknowledged, however, that researchers at the Middlebury Institute of International Studies’ Center on Terrorism, Extremism and Counterterrorism found GPT-2 can be taught to generate content supportive of various ideologies. Such automated propaganda could dramatically scale up the efforts of tech-savvy extremist groups. (You can experiment with a version of GPT-2 here.)

    Because it’s highly unlikely such advancements will be halted or that extremist propaganda efforts will be stopped, it’s time for us – the general public – to educate ourselves on these capabilities and the potential for misuse. Just as we’ve developed an ear and eye for airbrushed ad images, spam and telemarketing, we’ll need to do the same for content and how we consume it. As OpenAI itself says, “the public at large will need to become more skeptical of text they find online, just as the ‘deep fakes’ phenomenon calls for more skepticism about images.”

  • Performance-enhancing AI scandals will replace performance-enhancing drugs in professional sports. Researchers at Queensland University of Technology have created an AI system that can analyze a specific tennis player’s shot decisions to predict what they’re likely to do next. While the system is intended to enhance the spectator experience or support player training, it’s easy to imagine sports professionals receiving an alert during an actual game through an implantable or wearable device, signaling the next best action to take based on this intelligence.

    Unlike other major league controversies, like video-enabled sign stealing, the technology could be virtually undetectable – like a wearable embedded in a player’s leg that relays information through vibration.  The technique could extend beyond athletics to professional gambling or e-sports. It won’t be long before such ruses are uncovered, resulting in lost endorsements or prize money.

  • A business will get sued for (possibly unintended) unethical AI behavior. The potential for AI bias is much discussed, but very few companies have become adept at responsibly managing it. As AI is increasingly used to automating hiring processes, price quoting and other decision making, it’s inevitable that this lack of governance around bias will result in a lawsuit.

    Human resources personnel, for example, are trained to not ask about candidates’ family status, political orientation, religion, etc. However, AI systems can make such determinations by analyzing input, detecting correlations a human never would and making a decision based on the patterns it discerns. Similarly, credit card or insurance agents know not to assign rates and prices based on an applicant’s gender or race, but if not trained properly, an AI system (using facial recognition or natural language processing) might. When consumers begin encountering inherent bias, it won’t be long before the class action lawsuits begin.

    Most AI-driven unethical outcomes are not deliberate. Bias, however, is avoidable through technology measures, policy setting and good governance.  Whether through experience or trial and error, this will be the year businesses will need to up their game on AI ethics.

On a Positive Note

Now, to balance things out, let’s look at three events we can look forward to, gratis AI.

  • NASA will make a major announcement based on AI intelligence. Because it’s blazingly fast at detecting patterns from huge volumes of data and never gets fatigued in the task, AI can look at things in a way that people – and older space technologies – simply cannot. That’s why NASA has embraced the technology and has already made new space discoveries, like an eighth planet, with it.

    In 2017, the space agency fed an AI system with years of data collected by the Kepler Space Telescope and trained it to recognize the almost imperceptible change of brightness in a star’s light caused by an orbiting planet crossing in front of it.  As one astronomer says, it’s like looking for a needle in a haystack. “Once the AI learns the shape and appearance of a needle, it’s just a matter of examining each straw of hay in the stack, one by one, until it finds any that look like a needle.” Easy for AI; not so much for a human. Since then, two additional planets have been discovered using AI techniques.

    With its ability to quickly detect new relationships in huge volumes of data from space telescopes, radio systems, light waves and more, it’s inevitable that AI will – I predict this year – help identify a habitable planet for humans, without risking human lives in its discovery.

  • A new treatment will emerge for a major deadly disease. The use of AI in the medical field has been well documented, from breakthroughs in early detection of sepsis to increasing accuracy of breast cancer screenings to predicting suicide risk. By continuing to apply AI to not just detection but also prediction and prevention measures, medical researchers will likely make a discovery in the near term that will help practitioners reverse the impact of a previously intractable disease. The nature of AI is to identify possibilities that humans wouldn’t think to explore because we simply don’t see the patterns and correlations. Rather than “taking over” medical discoveries, AI will give doctors and researchers a much better chance of accurately detecting disease for individual patients and applying effective treatments and cures.

    As AI garners more attention for its impact on human lives, we’ll start to see the tables turn from “AI is dangerous” to “AI is a life-saver.” At Augusta Health in Virginia, an AI-driven sepsis predictor that uses vital signs and electronic health records data has saved 282 lives in two years by alerting nurses when factors for a patient rise above a certain threshold.

    This turnaround in AI’s reputation will shift our overall perspective on the use of the technology outside the medical field, such as in self-driving cars. Amid headlines on accidents and fatalities, we’ll also begin counting the lives saved by automated vehicles avoiding collisions or keeping interior temperatures safe when animals or children are inadvertently left in the car. Watch for the headline: “AI Saved My Life.”

  • Inclusivity and accessibility will boost quality of life for elderly and physically challenged populations. We’re already seeing AI-driven apps, wearables and other devices that translate text to Braille or speech, recommend walking directions to people with impaired sight or read text, recognize faces and identify objects for the user. With ever expanding AI augmentation, these developments will move from interesting experimentation to the mainstream in 2020, freeing people to live beyond physical constraints.

    Especially as they become baked into well-known products, and as price points decrease on specialized equipment, such capabilities will enable more independence and better quality of life for people who previously had no hope. In an increasingly aging population, this is essential – we’ll emerge from 2020 with less reason to fear the limitations brought on by age, illness or genetic tendencies.

With any of these predictions, it doesn’t take a leap of logic to see the business repercussions. As AI is used in increasingly mainstream but also imaginative ways, we’ll see the rise of performance-enhanced business people and performance-enhanced business processes. But rather than resulting in a scandal, the use of AI will simply become a key way of doing business in 2020.

For more AI predictions, see “20 Things to Expect from AI in 2020 and Beyond.” Learn more at our AI business practice

Bret Greenstein

Bret Greenstein

Bret is Senior Vice President and Global Head of Data at Cognizant, leading Cognizant’s Global Data Practice. He is focused on helping... Read more