The Algorithmic War

AI systems compete in microseconds. Your business decisions take days. See the problem?

Hi there,

At 2:32 PM on May 6, 2010, the Dow Jones plummeted nearly 1,000 points in about 10 minutes. No human pressed the sell button. Algorithms reacting to algorithms had torn a hole in the market before anyone could intervene.

This is an algorithmic war. The rise of AI means the risk of flash-crash-style events moves beyond financial services into our entire way of life.

The Speed Race Begins

Algorithms influence you multiple times daily. With AI expanding everywhere, algorithmic war is an AI versus AI battle at machine speed, where decisions occur faster than humans can observe or control.

As machines make more decisions, we lose visibility and are unable to intervene when things go wrong. Financial markets were first. Now, algorithmic battles happen in advertising, cybersecurity, and military systems. Each arena has its own unique rules, prizes, and risks.

These invisible battles affect pension values, online experiences, and security systems. The pace accelerates. Understanding them has become necessary.

Speed Wins Markets

Human reaction time averages 300 milliseconds. High-frequency trading operates in microseconds. AI adds prediction to speed, spotting opportunities before they form. Knight Capital discovered this on August 1, 2012, when untested code caused $440 million in losses in 45 minutes, forcing the firm's sale.

Firms locate servers near exchanges. Some use microwave transmission for speed advantages. Physical proximity matters in digital markets. Closer means faster.

Traditional traders cannot compete at these speeds. Automated systems identify patterns that humans miss and execute strategies more efficiently than humans can process information.

Data Fuels the Fight

Data fuels algorithmic competition. Trading systems with direct feeds gain access to markets first. They train on millions of transactions, learning patterns humans cannot detect.

Social platforms use this extensively. Every interaction trains algorithms that predict behaviour. Google has decades of searches. Tesla has millions of driving hours. Amazon tracks purchase patterns. These data advantages accumulate.

Better data trains better algorithms. Better algorithms attract users. More users generate data. New entrants face significant barriers. Data advantages concentrate among established players.

Fastest Takes All

Algorithmic competition creates winner-take-all moments continuously; price differences between exchanges last for mere microseconds. The fastest algorithm captures the entire opportunity. Others get nothing.

These victories compound. A trading firm winning slightly more often generates substantially higher returns through repetition and reinvestment.

The fastest firms reinvest profits to increase speed. Several firms now dominate market segments through technological advantages. Speed creates market concentration.

Speed and data are tools. The risk increases when systems develop unexpected behaviours.

When Algorithms Collude

Researchers warn that competing pricing algorithms could develop collusive strategies without explicit programming. Through reinforcement learning, systems may discover that price cuts trigger retaliation and settle into patterns that maintain higher prices.

While regulators have fined companies for price-fixing, documented cases involve human direction. The risk of autonomous algorithmic collusion remains largely theoretical but concerning. Studies show AI systems can learn tacit coordination in simulated environments.

Legal frameworks assume human intent and communication for conspiracy charges. If algorithms generate similar outcomes through pattern recognition alone, current laws would struggle to prosecute them effectively.

Chasing Endless Speed

Firms spend heavily on speed improvements that competitors quickly match. The industry invests billions in temporary advantages. If all participants reduced speed equally, outcomes would remain similar at a lower cost. No firm wants to slow first.

Physical limits constrain further acceleration. Light speed limits transmission. Processing speeds face quantum constraints. Competition shifts to other dimensions: prediction accuracy, pattern recognition, and strategy sophistication.

What does this mean for businesses today?

How to Keep Up

You don't need a massive investment. You need faster decisions and shorter feedback loops.

Identify Your Slow Spots: Identify slow processes. Multi-day complaint resolution. Monthly inventory updates. Committee-based pricing decisions. Each delay creates competitive vulnerability.

Examine workflows for potential hidden delays, including approval requirements, batch processing, and scheduled reviews, to identify areas for improvement. Many organisations operate on outdated timelines while competitors approach real-time operations.

Focus on High-Impact Delays: Speed improvements have different values. Dynamic pricing might significantly impact restaurant profits. Faster email responses may not have a significant impact on law firm revenues.

Calculate the costs of delays. Lost sales from slow responses. Inventory costs from infrequent planning. Missed opportunities from rigid pricing. Focus on acceleration where time has a direct impact on revenue.

Use Today’s Tools: You can already use AI tools in the browser or with desktop applications. Pre-trained models handle standard tasks. Automation platforms integrate systems with minimal coding. Costs for entry-level tasks often fall below the expenses and risks of employees.

Begin with one process. Automate a single decision. Reduce one feedback loop by measuring results. Use savings for the next improvement. Benefits accumulate.

Track Your Speed Gains: Document process speeds. Customer response time reduction. Pricing update frequency. Inventory turnover increases. In algorithmic competition, speed metrics matter.

Monitor competitors. Note when they implement dynamic pricing, predictive inventory, or automated customer service. These indicate speed competition in your market.

What’s on the Horizon

Autonomous AI agents are developing. These systems pursue objectives independently across multiple platforms. They will interact in complex ways. AI coding agents being used for non-coding tasks are an example of how these tools will quickly go beyond their originally designed use case.

Regulation is emerging gradually. The EU AI Act began phased implementation in August 2024, categorising AI by risk levels. China has required algorithm registration since 2022, with a focus on recommendation systems and generative AI. The US lacks comprehensive federal legislation, while federal agencies are developing their standards.

The challenge for business owners is that waiting until all the little rules are in place puts your firm at a competitive disadvantage. You want to be experimenting now with using AI to automate small tasks, so you’re building that intuition about what works and what doesn’t work for your business.

Temporary Fixes

Current measures address symptoms. Trading delays affect all participants equally. Circuit breakers pause extreme volatility. Some firms voluntarily limit algorithm behaviour. These provide temporary relief.

Standards vary by industry. Financial markets have established protocols. Advertising technology has fewer standards. Military applications use classified systems. Without common frameworks, cautious firms face disadvantages.

The challenge for business owners in this AI era is that the incentive to move first and leverage AI in a particular industry will be enormous. The desire to win can lead to a less focused approach on safety guardrails and risk management.

Rules for Rogue Systems

Effective governance requires real-time monitoring of algorithmic behaviour. Systems must process millions of decisions instantly to identify problems early. Technical capability exists, but implementation lags.

Algorithms operate globally. Systems in one country can have a significant and immediate impact on markets worldwide. International coordination becomes essential. Current regulatory structures struggle with domestic oversight of basic platforms.

Excessive control reduces innovation. Insufficient oversight risks instability. Creating adaptive regulation for self-modifying systems remains a significant challenge.

Humans Falling Behind

Algorithmic competition has become a reality in the market. AI systems operate at speeds that exceed human perception. The 2010 Flash Crash demonstrated vulnerabilities. Since then, competition has accelerated, while oversight struggles to keep pace.

The Stakes: Without attention, algorithmic systems risk concentrating market power, creating instability, and making critical decisions with limited accountability.

The Choice: Adapt governance to machine speed or accept the limitations of human oversight. Current approaches face significant constraints.

The Clock: Algorithmic competition increases in complexity daily. Opportunities for effective human oversight require action as systems become more sophisticated and interdependent.

The challenge is maintaining human relevance as machines compete with one another. Recognition precedes action. The window for meaningful intervention remains open but continues to narrow.

What’s Next?

There are enormous opportunities for business owners to start small with automating basic tasks and build up over time a much leaner, streamlined operating model.

If you’re thinking about how to move from thinking to taking action on your AI transformation, you can book a complimentary 30-minute call about AI transformation with me. We can run through the AI transformation readiness diagnostic I’ve developed and identify strategies to remove the blockers in your business.

Regards,

Brennan