Fair Observer’s operations chief of staff Cheyenne Torres and entrepreneur Dirk Lueth, co-founder and co-CEO of Upland, examine how artificial intelligence is reshaping economies, employment and the media. Lueth explains why today’s generative AI differs from traditional software and traces its development from an academic concept to a widely accessible technology. Although he acknowledges the risks of job displacement, concentrated corporate power and excessive investment, he expects AI to increase productivity and create new opportunities. He believes that human judgment, trusted institutions and energy infrastructure will become increasingly valuable as AI grows more powerful.

From rigid rules to statistical models

Torres begins by asking what AI actually means. Lueth distinguishes it from traditional software, which relied on programmers writing deterministic rules. Large language models instead analyze vast amounts of human-created material and identify statistical patterns in language. They can generate useful responses across fields ranging from medicine to law, but Lueth stresses that they have no consciousness or human understanding.

He also distinguishes chatbots from AI agents. A chatbot usually responds to an individual request, much like a search engine. An agent can perform a continuing assignment, complete multiple steps and act more like a virtual worker.

AI itself is not new. Researchers coined the term at Dartmouth College in 1956, but Lueth describes the following decades as a period of overpromising. Technical breakthroughs eventually brought AI closer to public expectations. AlexNet advanced image recognition in 2012, Google researchers introduced the transformer architecture in 2017 and ChatGPT made generative AI accessible to the public in 2022.

Expertise gets cheaper

Lueth considers AI revolutionary because it spreads knowledge and reduces the cost of accessing expertise. People can now ask an AI system for help with household repairs or basic financial and legal questions instead of immediately consulting a professional. Scientists and physicians can also use the technology to accelerate research or analyze medical images.

Businesses may experience equally profound changes. Lueth says AI allows smaller organizations to compete with larger companies because employees can accomplish more with fewer resources. He estimates that Upland, a virtual property game that combines elements of Monopoly and The Sims, operates with roughly 20–30% of the staff it might previously have required to build and operate such a game.

Yet the revolution carries significant costs. AI systems consume large quantities of energy while much of the technology remains concentrated among a few powerful companies. Lueth is also concerned about entry-level employees. AI can perform many of the routine research and administrative tasks through which young workers traditionally acquired experience, potentially weakening the first steps of the career ladder.

Standardized jobs under pressure

Torres asks which workers are most vulnerable. Lueth identifies jobs involving predictable and standardized processes, including paralegal work, call centers and business process outsourcing. AI can already handle many customer service requests instantly, offering consumers an attractive alternative to waiting for a human representative.

However, Lueth cautions against attributing every recent layoff to AI. Some companies expanded excessively during the Covid-19 pandemic and may now invoke the technology to explain ordinary restructuring. He calls this practice “AI washing.”

Lueth nevertheless accepts that some workers will suffer as companies automate existing processes. The long-term outcome will depend partly on whether displaced employees can learn new skills and move into roles created by AI. He compares the transition to earlier technological changes that eliminated some agricultural jobs while producing entirely new industries. Web designers and data scientists, for example, became common professions only after the Internet created demand for them.

Although future occupations are difficult to predict, Lueth expects workers equipped with AI to become far more productive. He therefore believes the overall economic effect will be positive even if the transition proves painful for particular workers and communities.

AI investment may be excessive

Lueth separates the supposed AI bubble into usage and financial components. He sees no usage bubble because individuals and businesses continue finding practical applications for the technology. More capable agents and intelligent robots will likely reinforce demand rather than cause it to disappear.

The financial picture is less reassuring. Investors are directing enormous amounts of capital toward a small group of companies, while startups may struggle to raise money unless they present themselves as AI businesses. This concentration could deprive other productive sectors of investment.

Lueth also points to circular financing arrangements. A chipmaker might invest in an AI company that subsequently spends the money on that chipmaker’s products. Such transactions can inflate apparent demand and obscure whether the industry is generating sustainable value.

Regulation could protect citizens from fraud and other malicious uses of AI. However, Lueth warns that complicated rules may produce regulatory capture by imposing costs that large technology companies can absorb but smaller competitors cannot. He favors basic government protections combined with meaningful industry self-regulation.

Trust, taste and energy define the future

Lueth imagines AI gradually becoming an invisible commodity like electricity or the Internet. Phones and computer screens may give way to conversational interfaces, translation earpieces and other tools that allow people to interact naturally with machines.

Employment could also change fundamentally. Applicants may eventually present portfolios of specialized agents that write copy, produce videos or conduct research on their behalf. Some software itself may become disposable as individuals create applications for temporary needs and abandon them afterward. As technical expertise becomes abundant, Lueth says that taste, trust and accountability will become scarce and valuable human qualities.

These qualities will be particularly important for the media. Realistic synthetic audio and video will make it increasingly difficult to determine whether an event actually occurred. Lueth believes trusted publications will therefore become essential sources of verification.

Finally, he predicts that geopolitical competition over AI will ultimately depend less on which country creates the best model than on which can produce enough energy to operate the necessary infrastructure. Nations and regions capable of expanding reliable energy supplies will be best positioned to convert AI into productivity, scientific progress and economic power.

[Lee Thompson-Kolar edited this piece.]

The views expressed in this article/video are the author’s own and do not necessarily reflect Fair Observer’s editorial policy.