By Eric A. Coop
When one hears news about artificial Intelligence (AI) or machine learning (ML), one understandably thinks of programs such as OpenAI’s Chat Generative Pre-Trained Transformer, better known as ChatGPT; but AI is much more than ChatGPT. AI and ML are poised to affect every aspect of our lives, promising, despite serious challenges, to positively impact communications and spectrum management, among other fields.
Understanding the Threats of AI
AI technology which “scrapes” existing content from across the Internet has raised concerns across fields. Academia is split between rejecting it as a form of cheating and embracing it as a new learning tool worthy of exploration. Writers and other producers of monetized content often see AI as a threat to their livelihoods, exemplified by two lawsuits that actor, comedian, and author Sarah Silverman recently filed against Meta and OpenAI.
AI also raises significant social, existential, and safety concerns. Writing in the Financial Times on July 24, 2023, Secretary of State Antony Blinken and Secretary of Commerce Gina Raimondo acknowledged the risks of unmitigated AI in perpetuating “false information, reinforcing bias and discrimination, being misused for repressive or destabilizing purposes,” proliferating bioweapon design knowledge, or enabling cyber-attacks. We can expect to see AI threaten the 2024 presidential election, with bad actors breathing new life into familiar disinformation tactics. AI has already been used to generate “deep fakes,” creating photorealistic images and videos, and convincingly human voice and text.
In recent years, scammers have employed AI to identify individuals through their phone numbers, even in anonymous databases. At a July 13 forum on Opportunities and Challenges of Artificial Intelligence for Communications Networks and Consumers held by the Federal Communications Commission (FCC) and National Science Foundation (NSF), Dr. Alisa Valentin, Senior Director, Technology and Communications Policy, National Urban League, highlighted the predatory nature of robocalls and the need for all communities to have access to fair and reasonable communication prices. She emphasized the importance of incorporating anti-discriminatory provisions into existing and proposed telecommunications rules as AI becomes more prevalent.
Equally concerning is the sophistication of deep fake robocalls. Thieves are now using AI and ML to impersonate relatives and authority figures. While such scams have been around for years, advancements in voice cloning are making these robocalls even more convincing. Ditto for deep fake videos of famous people saying and doing things they did not. Just ten years ago, scam calls were clunky and edited video of, for example, President Obama kicking a door were funny but not particularly convincing. It is a different world now.
To combat these threats, new consumer protection tools are using AI and ML to intercept and screen incoming calls for spam and have been shown to be more effective than legacy blocklists. One new consumer application, Jolly Roger, employs ChatGPT and voice modulation to “keep telemarketers on the line indefinitely.” The author has used both Jolly Roger and the AI tool Hiya to screen and block calls.
While dialogue on the potential adverse ramifications of AI is dominating much of the media coverage, AI also promises innovations and benefits in several areas. Indeed, recent commentary by Secretary of State Antony Blinken, Secretary of Commerce Gina Raimondo, and FCC Chair Jessica Rosenworcel separately lightened the dark picture of national security concerns with rosier promises of spurring innovation and overcoming some of the greatest social challenges we face. An AI Bill of Rights and AI Risk Management Framework are under development to mitigate the destabilizing risks.
It’s Not All Bad News
Recent developments illustrate that telecommunications regulators are taking AI seriously. During the July 13 FCC-NSF forum, a diverse range of stakeholders gathered at FCC headquarters to discuss the promise and challenge of AI in the telecommunications and technology sectors. Topics ranged from AI’s transformative potential to perfect network traffic to its ability to enhance resiliency through self-healing networks.
At its August 3 Open Commission Meeting, the FCC continued its exploration of AI, unanimously approving a Notice of Inquiry (NOI) to “initiate a technical inquiry into how to obtain more sophisticated knowledge of real-time non-Federal spectrum usage” using “new data sources, methods, and technologies such as artificial intelligence and machine learning in an increasingly congested radiofrequency environment.”
Chair Rosenworcel expressed optimism regarding the possible application of AI and machine learning in (1) supporting real-time dynamic spectrum sharing and decision-making to increase spectrum efficiency while decreasing interference, (2) autonomously troubleshooting and repairing networks to enhance network resiliency, and (3) protecting consumers by expanding and refining analytics tools to block robocalls and robotexts.
Indeed, AI-powered real-time dynamic spectrum sharing holds promise in that it is expected to automate spectrum management to mitigate interference concerns far more efficiently. Spectrum bands which are, for instance, reserved for military applications could become partially available to other users. The Pentagon and commercial 5G providers are currently struggling, for example, with defining an efficient approach in the 3.1-3.45 GHz band. A U.S. Navy warship operating on the high seas might need exclusive use of these frequencies while a 5G network in the Midwest could use the same frequencies without causing a problem. In coastal regions, AI-powered dynamic spectrum sharing could open more of these contested frequencies to commercial users with carefully defined and managed limits to protect incumbents while maximizing value for new entrants.
Network resilience is likewise becoming increasingly important. The perpetual struggle against bad actors who look to hack networks has been compared to a “cat and mouse” game. As service providers incorporate advances in AI and ML, bad actors are likewise incorporating their own advances. There are, of course, other threats to network resilience, and applying AI and ML tools to network management is expected to lead to autonomous troubleshooting and self-healing at faster rates than even the best Information Technology personnel can facilitate.
To address the social and economic threats of AI to traditionally marginalized communities, the industry is fighting fire with fire. Developing technologies that will employ AI and ML to combat targeted robocalls, robotexts, and disinformation campaigns include the Secure Telephone Identity Revisited (STIR) and Signature-based Handling of Asserted Information Using toKENs (SHAKEN) standards. STIR/SHAKEN will incorporate caller ID signature validation on interconnected phone networks as well as digitally validating the phone call handoffs through complex networks to allow the phone company of the consumer receiving the call to verify that a call is in fact from the number displayed on Caller ID.
Finally, the latest National Artificial Intelligence Research and Development Strategic Plan (updated in May 2023) contains nine strategies. The previous eight strategies contained in the 2019 update underscored the importance of broader concepts such as long-term investment, ethical/legal/social implications, and trustworthiness. The ninth strategy seeks “to underscore a principled and coordinated approach to international collaboration in AI research.” This strategy is intended to restore American leadership in AI development by setting the standards & frameworks and promoting AI as a tool for global benefit.
The Prognosis
The future of AI and ML is a bright one. These rapidly-advancing technologies will protect consumers, enable efficient real-time spectrum management/deconfliction in highly congested spectrum, and become more-embedded in all aspects of our lives. In a previous blog post, the author painted a rosy picture of the Internet of Things (IoT) and the promises it holds. AI and ML will enable the IoT, our future smart cities, 6G, the realized gains of the Broadband Equity Access and Deployment Program (BEAD), and the mitigation of historic inequities affecting marginalized communities. The criminals will continue to find ways to exploit AI and ML against vulnerable people, but the industry is far better poised to stay ahead of the threat.
The views and opinions expressed in this blog post are those of the writer and do not necessarily reflect the views or positions of any entities they represent.