What falling price of TVs should teach us about the future of AI and work: David Richards
In the early 1970s, my parents rented our colour television. So did almost everyone else we knew. You did not casually buy a TV. You signed a rental agreement. That enormous wooden box sat in the corner of the living room like a prized possession on loan.
A decent set cost around £250; more than £4,500 in today’s money. Imagine spending that on three channels, heavy buttons, and no remote.
In 2026, you can buy a vastly superior 4K flat screen for roughly the same nominal £250. Fifty years of inflation have passed, yet the sticker price looks eerily similar.
A worker on the National Living Wage can earn that television in roughly 24 hours today.
The same worker would need decades to buy the average semi-detached house that a 1970s family purchased for three times their annual salary.
The difference is productivity. Televisions became cheap because we learned to produce them at extraordinary scale. Automation and software-controlled manufacturing drove down the cost of production. We now get more output from fewer inputs.
Artificial intelligence is about to do the same thing to intelligence itself.
For centuries, human judgment and expertise were scarce. You paid for a lawyer’s time because legal reasoning was rare. You hired engineers because the ability to build complex systems was specialised. AI is rapidly driving down the marginal cost of high-quality analysis, drafting, coding and diagnosis. What robotics did for manufacturing, AI is doing for cognition.
That creates understandable anxiety. If intelligence becomes abundant, what happens to those who sell it?
We have long assumed that full employment is the foundation of a healthy economy. Fewer traditional jobs means something has gone wrong. That assumption may not survive the AI era.
Project forward to 2036. AI is embedded in operating systems, integrated into every company and running behind public services. Every individual has a suite of agents handling research, planning, design and administration.
A start-up in Sheffield can access legal drafting, financial modelling and marketing strategy at near-zero cost. A single founder can do what once required a team of 20.
Routine bureaucracy is largely automated. Medical triage, tax processing and benefits administration run continuously. Fewer hours of routine cognitive labour are required to keep society functioning.
In that world, we may need a new economic baseline.
Today’s unemployment support is designed as a safety net. It is conditional, means-tested and often carries stigma. It assumes work is available and that the recipient is temporarily outside the labour market.
A universal minimum income would acknowledge something different. If AI-driven productivity dramatically lowers the cost of goods and services, society can afford to guarantee a basic income floor, not as charity, but as a dividend of automation. This is not about paying people to do nothing. It is about recognising that when machines generate extraordinary economic value, citizens should share in the gains simply by being part of that society.
The distinction matters. Unemployment benefit implies personal shortfall. A universal minimum income would reflect structural abundance.
Return to that rented television. What once required a rental contract and a family debate can now be bought with a week’s wages. Technology quietly raised the baseline of what was normal.
AI is moving far faster than television manufacturing ever did. The productivity gains that took five decades to transform electronics may take ten years to transform intelligence.
If that happens, the real challenge will not be preserving every traditional job. It will be building an economy confident enough to share abundance, even when work no longer defines us in the way it once did.
David Richards MBE is managing partner of Yorkshire AI Labs