Artificial intelligence Part 8: specific industry impacts - retail sector

ChatGPT image

The retail industry has managed several significant disruptions, particularly with technology, over many decades. Retail in this definition, spans both the physical and digital environments with multiple different segments and collectively has the highest overall level of employment. The impact of AI occurs differently across the segments as discussed below -

E-commerce and digital retail

Disruption caused by AI is most advanced in this segment of retail -

• Merchanidising and product curation: AI is already used online to determine what products are shown, to which potential purchasers at what price and in which order. The human merchandiser role that once required deep product knowledge and market intuition is being replaced by AI. The result is buying teams are being reduced in headcount.
• Pricing and promotions: AI can operate dynamic pricing in real timer across millions of stock keeping units (SKUs) based on demand, competitor pricing and inventory levels. The pricing analyst role has been largely automated.
• Demand forecasting and inventory management: this is a large team function traditionally in retail requiring significant human judgement (which sometimes was unsuccessful with some product lines), however now AI is a dominant force. Inventory planners are increasingly deployed to exception management only.
• Product descriptions, copywriting and promotional content: once the function of large copy teams for catalogues, AI now generates product listings at scale for example, at Amazon, ACOS and Zalando. 
• Customer service: Tier 1 customer service in e-commerce had become largely automated with chatbots and AI agents now capable of handling item returns, order queries, and complaints at scale. Human agents are increasingly handling ecalated issues only.
• Searches and recommendations: the entire discovery layer of e-commerce is AI-driven thus reducing the need for manual curation teams.

Physical Retail - Shops, boutiques bricks and mortar stores

The physical retail segment has a much stronger protective layer being the embodied, social experience of shopping and the overall 'retail therapy' of personal service, expertise and the physical handling of possible items for purchase. Luxury retail, specialist retailers together with experience-led formats have led to an investment in more knowledgeable human staff. Consumers expect such interaction, expertise and better service. Retailers whom do not offer higher level services for more expensive products often find public criticism including negative consumer ratings online are the result. Nonetheless there is AI intrusion and pressure in this segment -

• Checkout and payment: the increasing use of self-checkout from stores and the replacement of human checkout counters has been the proverbial thin-edge-of-the-wedge for the past two decades. For example, Amazon Go and similar checkout models eliminate checkout operator roles entirely. This process started in gocery stores but has expanded across a large part of the retail industry however not without some caveats - staff are needed to supervise and monitor self-checkout store sections, increased surveillance for fraud and theft has become essential and human operator checkout aisles have needed to be retained in smaller form. 
• Stock management and replenishment: computer operated vision and AI inventory systems can detect shelf gaps and product depletion thus triggering replenishment action without human stock checkers. Warehouse stock picking has increasingly become robotic.
• Loss prevention: AI camera systems are increasindly replacing human loss prevention officers for monitoring and surveillance in-store. Human response to any alerts however remains essential.
• In-store customer assistance: customer assistance where genuinely helpful product knowledge  roles are needed means there positions are much more protected from AI encroachment. Generic customer greeting roles are much less so.
• Visual merchandising: AI tools can optimise store layouts using foot traffic data and correction with sales information. This reduces the creative and analytical work of visual merchandisers.

Grocery and Fast Moving Consumer Goods (FMCG) Retail

Grocery and fast moving consumer goods is another segment when AI operates removing more of the operational back office roles. For now it cannot do the physical work of storage, presentation and despatch however systems operation is custom made for AI.

• Category management: the analytical heavy lifting that category planners (planogram design, range rationalisation, promotion assessment and evaluation) are all in the process of being automated.
• Supply chain coordination: AI optimises routing, supplier ordering and waste reduction. The impact is the reduction of operational planning team headcount.
• Fresh food management: AI waste reduction tools are already deployed in major supermarkets. This is a direct threat to manual stock management roles but not to actual physical movement of materials.

A further post in this blog will cover other aspects of the retail industry where AI is impacting marketing, advertising and retail fashion.

Climate change - the option of algae instead of biofuels ?

 

Many biofuels haven’t panned out. Could algae make the clean diesel and aviation fuel Australia needs?

Peter Ralph, University of Technology Sydney; Alexandra Thomson, University of Technology Sydney, and Martin Lloyd, University of Technology Sydney

Diesel is critical to Australia. Any supply disruption has immediate and widespread consequences, given Australia imports almost 80% of its liquid fuels. As the energy shocks of the Iran war ripple out, Australia’s leaders have scrambled to shore up supplies of fuel – especially diesel and aviation fuel.

Disruptions to fuel supplies have happened before, such as in 2008 and 2022. This disruption won’t be the last.

What should policymakers do? One option is to ramp up local production of biofuels made not from crude oil but from natural oils such as canola, animal fats – or algae.

As algae researchers, we believe these humble organisms are worth exploring. Making biodiesel and sustainable aviation fuel from these fast growing organisms can be done with much less land than other crops. Technological advances mean the fuel could scale up.

Many biofuels come with trade-offs

Biofuels have gained traction worldwide as efforts to reduce dependence on fossil fuels and meet climate targets accelerate.

The Australian biofuel sector is relatively small. Farmers exported about 6 million tonnes of canola in 2023–24 to be turned into biofuels overseas.

The Australian government last year announced A$1.1 billion in incentives to boost low-carbon fuels such as biofuels.

Biofuels from corn, soybean, canola and palm oil have boosted fuel security in some nations. Brazil produces 22% of its own transport fuel from biofuels, while biofuels account for 6% of the fuel used in the United States.

The problem is, biofuels often come at an environmental cost. A third of all US corn is used to make ethanol for fuel.

What’s so good about algae?

The type we’re interested in are microalgae, single-celled organisms, not macroalgae such as kelp and other types of seaweed.

These small organisms can grow exceptionally rapidly and hold high concentrations of oils. Many microalgae species can double their weight every day. Nannochloropsis and Chlorella are the two main types used to make oil.

Traditionally, algae was grown in large, shallow outdoor pools called “raceways”. They’re now increasingly grown in high-efficiency algae bioreactors.

Algae can be processed using proven technologies such as hydrothermal liquefaction to produce biodiesel able to be used in existing trucks and machinery. It can also produce sustainable aviation fuel.

Compared to crop-based biofuels, algae has several advantages. It doesn’t compete with food production and it can be grown on non-arable land or in industrial facilities. Some species can grow in saltwater or even treat wastewater while using it for growth. If algal facilities are located near heavy industry, carbon emissions can be captured and used for algal growth in a form of carbon storage.

Algal fuels needs much less land than conventional biofuels. A hectare of algae can yield more than 58,000 litres of oil per year. By contrast, a hectare of corn produces just 172 litres.

What are the barriers?

Interest in algal fuel dates back many decades. Oil shocks in the 1970s and 1990s drove significant research into algae-based fuels. But when oil prices fell, algal biofuels were no longer cost-competitive.

Since the 1990s, technologies have matured and policy settings become more favourable. Efforts to reduce fossil fuel use have put an implicit or explicit price on carbon. Mandates to increase output of sustainable aviation fuel are emerging in the European Union.

Fossil fuel price shocks in 2022 and 2026 have nudged authorities to seriously explore alternatives. Sovereign fuel security has become a strategic priority. Both the United Arab Emirates and the US are exploring algal fuels as a long-term strategic asset.

Algae for Australia?

Australia would be well placed to explore the potential of algal fuels. It has plenty of non-arable land, abundant sunlight and some of the world’s best algae research capabilities. Plus, it depends very heavily on imported diesel and aviation fuel.

Our research group and many others have been systematically working to overcome previous limitations of algal biofuels. We now know how to produce high-quality algal fuels and scale up production at costs low enough to challenge fuels derived from crude oil.

The first step would be to invest in pilot projects to prove the technology can work at scale under real-world conditions. Overseas, similar pilots have been set up next to industry to test the use of carbon capture, or alongside research partners.

If this is successful, the next step would be to build facilities in regional locations where fossil diesel is in demand and expensive to transport – and where algae can offer a dual benefit by treating wastewater or capturing carbon.

Over time, the versatile technology could be expanded, as algae can produce not only biodiesel but also other useful products such as edible protein for animal feed and biochar, highly porous charcoal able to soak up pollutants such as heavy metals.

Researchers have been working to boost yields and scale up oil production from algae. mayaluana, CC BY-NC-ND

Algae deserves our attention

Many previous efforts to scale up biofuels have run into problems over environmental impact or cost.

It’s important to be sceptical of claims of the next big thing. But it’s also important not to overlook the potential of humble technologies such as making fuel from algae.

As leaders look for ways to bolster fuel security, algae deserves a closer look.

Peter Ralph, Distinguished Professor of Marine Biology and Executive Director of the Climate Change Cluster, University of Technology Sydney; Alexandra Thomson, Industry Engagement Manager, Climate Change Cluster, University of Technology Sydney, and Martin Lloyd, Strategic Lead, Research Translation, University of Technology Sydney

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Artificial intelligence Part 7: specific industry impacts - corporate governance and risk management

ChatGPT image

The impact of artificial intelligence on governance and risk management is nuanced. There are some tasks that AI can do, but other functions must be undertaken by human beings. As a general description of work functions, governance teams usually produce documentation such as board briefing papers, governance reports, policy updates, compliance documentation and meeting summariess. These are all processes open to AI. This posting will provide an indication of changes coming to roles in corporate governance and risk management due to AI. 

Corporate Governance

• Board reporting and briefing papers preparation: AI can synthesise management informaiton, financial data and identify risks into baord-ready material reducing the administrative teams that currently do this work.
• Corporate secretariat functions: the various documents produced such as records of minutes, compliance tracking, regulatory filings are all highly structured functions that are automatable and suited to AI.
• ESG reporting: Environmental, Social and Governance (ESG) reporting requirements have increased in recent years creating an additional compliance burden. AI can offset the increase through tools that aggregate and report sustainability data.

Risk Management

• Credit risk modelling: Credit risk modelling is a heavily quantitative task suited to AI thus reducing the analyst layers that do the manual model running and reporting. 
• Operational risk assessment: risk assessment identifies and quantifies risk across various buinsess processes. The assessment often produces a risk assessment table which rates and grades the risk including mitigation method. These processes are suited to be augmented by AI at the data gathering and first-analysis stage.
• Market risk monitoring: real-time AI surveillance is replacing/reducing some human monitor headcounts for this task.

Compliance and regulation

• KYC (Know Your Customer) and Anti-Money Laundering (AML): AI is transforming these tasks from labour-intensive manual processes to AI-monitored exception-handling workflows. The impact is the reduction in the size of large compliance teams. 
• Regulatory change management: the tracking and interpreting of new regulations can be augmented by AI for monitoring and summarising only. The interpretation of ambiguous regulatory language remains an essential human function not AI.
• Audit: the largest 4 accounting firms are deploying AI to analyse entire transation populations rather than the traditional scope of only sampling. This change provides a better quality audit but it requires fewer junior auditors.

Prudential Regulation (Central Banks and Regulators)

• Supervisory data anlysis: regulators such as APRA (in Australia), FCA (US) and Central Banks are using AI to monitor systemic risk across institutional data.
• Examination and inspection of institutions: the inspection teams face efficiency improvements and possibly headcount pressure.
• Policy and rule-making: roles in these functions are more protected however the judgement and accountability requires is high for human decision making.

In terms of employment, senior risk professionals are still needed for a range of actions such as - determining acceptable risk levels, advising executives on risk management, assessing complex of emerging risks, balancing regulatory, financial and reputational considerations. 

Physical inspection teams are still needed and remain human-based for purposes such as: construction safety inspections, infrastructure maintenance checks, environmental site visits, equipment integrity inspections. 

In the future, the workforce changes are more likely to be a reduction in the entry-level and junior operational roles with a structure as shown -

Before
Chief risk/governance officer

Senior specialists

Large analyst and reporting teams

After

Chief risk/governance officer

Senior specialists (fewer)

Small team supervising AI monitoring systems

The AI systems in use include: Microsoft Copilot, IBM Watson, Palantir Technologies and SAS Institute.

AI example - star series - black hole - animation using html code

 

Health - Hantavirus explanation

 

What is hantavirus, the disease that has killed 3 cruise ship passengers?

AFP/Getty

Thomas Jeffries, Western Sydney University

Three people have died after a suspected outbreak of hantavirus on a cruise ship in the middle of the Atlantic ocean. At least one other passenger is in intensive care in South Africa.

The World Health Organization announced the deaths in a social media statement on Monday, along with one confirmed case of the rare disease. Authorities are investigating another five suspected cases among passengers travelling on the MV Hondius.

So, what is hantavirus? And why can it be so deadly?

As the investigation unfolds, here’s what we know.

What is hantavirus?

Hantavirus is a rare but severe respiratory illness that can cause severe bleeding, fever and even death.

The virus is spread by rodents, such as mice and rats, mainly through the urine and droppings of infected animals.

Hantavirus does not typically spread from person to person. However, in rare cases it may spread between people.

Globally, there are an estimated 150,000 to 200,000 cases of hantavirus each year.

It is less contagious than airborne viruses such as COVID and influenza, as it typically does not spread from person to person.

What makes it so deadly?

There are two main types of hantavirus, each with different symptoms.

Hantavirus pulmonary syndrome, which affects the lungs, is mainly found in the United States. If a person becomes infected with this type of hantavirus, within days they will likely experience coughing and shortness of breath.

As the illness progresses, they can develop symptoms such as fatigue, fever and muscle aches. They may also get headaches, dizziness, nausea, vomiting and abdominal pain. This is the most deadly kind of hantavirus. Tragically, about 38% of people who develop these symptoms die from the disease.

Hemorrhagic fever with renal syndrome is mainly found in Europe and Asia, but the strain known as the Seoul virus has spread around the world. This form of hantavirus mainly affects the kidneys.

People usually develop symptoms within two weeks of being exposed to this virus. Early symptoms include severe headaches, abdominal pain, nausea and blurred vision. More advanced symptoms include low blood pressure, internal bleeding and even acute kidney failure. This disease can be caused by different viruses and some are more deadly than others, meaning between 1% and 15% of cases can be fatal.

Unfortunately, there is no specific treatment or cure for either type of hantavirus. However, early medical treatment may increase a person’s chance of survival. This can include using respirators, oxygen therapy and dialysis.

Authorities are still investigating which type of hantavirus the passengers were exposed to.

How did it get on a cruise ship?

In a closed environment such as a cruise ship, there are two possible ways passengers could have contracted hantavirus.

One is being exposed to the virus while on a shore excursion.

The other possibility is that rodents may have entered the ship on cargo, and then spread the disease to passengers through their infected urine or droppings. Other factors such as hygiene standards and food storage practices may have caused the infection to spread more quickly.

To contain this suspected outbreak, authorities must first ensure any rodents are safely contained and removed from the ship. They should then monitor all passengers for hantavirus symptoms. The virus is diagnosed with a PCR test, similar to those used to diagnose viruses such as COVID.

Given there is no specific treatment for the disease, authorities must help any infected passengers manage their symptoms. This involves checking that they are breathing normally and their kidneys are functioning properly.

So, how worried should we be?

Although alarming, cases of hantavirus remain are extremely rare. But it can look similar to other respiratory illness, so you should always get symptoms checked. If you’ve been in regions where the virus is found and experience shortness of breath, fever or any other flu-like symptoms, see your GP.

Thomas Jeffries, Senior Lecturer in Microbiology, Western Sydney University

This article is republished from The Conversation under a Creative Commons license. Read the original article.

AI example - star series - neutron star - Pulsar animation using html code