Recent allegations by the creators of GIPPR, a conservative chatbot based on ChatGPT technology, that OpenAI shut down their project heighten concerns that artificial intelligence technology will be used to limit certain viewpoints. GIPPR’s developers allege that OpenAI said their bot violated policies regarding “deceptive activity and coordinated inauthentic behavior.” Testing, however, showed that the bot repetitively identified its partisan viewpoint with statements such as “as a right conservative AI, I” and “as a conservative AI, I believe.”

Beyond the dispute between GIPPR and OpenAI, concerns about the politics of technology are not new. For example, German/American rocket scientist Wernher von Braun was the subject of recurrent criticism regarding his willingness to first develop a rocket for the World War II German military and then switch to rocket development for the U.S. military and NASA .

The implications of artificial intelligence development are no less significant than those of the Saturn V or V-2 rocket. Some have, with unnecessary hyperbole, contended that improperly managed AI could lead to the ultimate calamity of human extinction . More realistically, AI has helped the disabled , improved medicine , made real estate agents more efficient , and helped organizations recover from cyberattacks . In the future, it is poised to aid humans in numerous areas ranging from improving education to drive-thru restaurant ordering .

But given the many studies that show scientists’ political inclinations (based on donation analysis) lean toward the Left , it’s easy to be concerned that AI may not represent Republican, libertarian, or numerous other views. With its potentially pervasive future role in society, biased AI is inherently problematic.

And it’s not just partisan bias that should be concerning. AI can also have racial/ethnic , gender, geographical, and numerous other biases — potentially due to the underlying technology, but in many cases, due to the biases of a system developer or the data that the AI is built or trained with.

Regulation and licensing of AI technologies have been floated as a means to prevent demographic biases from affecting AI operations. This can indeed be effective as many are already regulated, and existing laws may already cover problematic behavior, such as discrimination. However, application regulation doesn’t allow government control of the underlying algorithms or their speech. Both the programmer's code and the application's recommendations may be protected expression, covered by the First Amendment.

There is, similarly, no current basis for the government to require, under most circumstances, that algorithm developers make their technology available for others to use. However, this is not inconceivable. Models used for phone company regulation , such as making parts of the phone network available to all carriers, may provide some insight into the implications of doing so.

Fortunately, those (including government agencies) who wish to shape AI development have a strong tool: financial incentives. Agencies can fund technology development with particular goals, using grants and contracts, or they can create incentive programs (modeled after NASA’s Centennial Challenges program, for example) to encourage developers to undertake these efforts independently. Companies and entrepreneurs can invest in or start firms to develop technologies or contract with others to do so.

Similarly, political parties may benefit from funding technology development aligned with their goals and objectives. This provides a mechanism for getting technologies desired by agencies, firms, or individuals developed. It also increases the overall level of resources available for AI, further expanding technology development.

The public interest sparked by ChatGPT and similar technologies shows the challenge posed by new AI technologies. It is crucial for stakeholders, from individual developers to large corporations and government agencies, to appreciate the significance of diverse perspectives in the development and implementation of AI technologies. It's only through collective efforts that we can hope to harness the full potential of AI.

Jeremy Straub is the director of the North Dakota State University’s Institute for Cyber Security Education and Research, an NDSU Challey Institute faculty fellow, and an assistant professor in the NDSU Computer Science Department.